Remote sensing image security retrieval method and system in cloud environment

Abstract

The disclosure discloses a remote sensing image security retrieval method and a remote sensing image security retrieval system in cloud environment. According to the present disclosure, a CNN model is applied to extract features, then dimension reduction is performed according to spectral hashing with spectral rotation, and clustering is performed through K-means++. A secure encryption searchable index is generated from the correction tuple of Asymmetric Scalar-Product-Preserving Encryption (ASPE). Image encryption is performed after creating watermark embedding space through prediction error marking and pixel rearrangement. The copyright information and the user information are flexibly embedded and extracted in the process for copyright authentication and distribution data tracking. In addition, the authenticity of the search result is verified by constructing the Merkle tree.

What is claimed is: 
     
         1 . A remote sensing image security retrieval method in cloud environment, comprising:
 S1, by the image owner, utilizing a pre-trained CNN model to extract features from the outsourced image, generating an encrypted searchable index based on spectral hashing with spectral rotation for dimensionality reduction and ASPE, carrying out prediction error marking and pixel rearrangement to create space for watermark embedding, encrypting the image set and constructing a Merkle tree, transmitting the encrypted image set and encrypted searchable index to a cloud server, and sending the Merkle tree to the certification authority;   S2, by the query user, utilizing the pre-trained CNN model to extract features from the query image, generating a query trapdoor based on spectral hashing with spectral rotation for dimensionality reduction and ASPE, transmitting the query trapdoor to the cloud server, and subsequently verifying the authenticity of the search result through the certification authority;   S3, storing, by the cloud server, outsourced images and the encrypted searchable index of the image owner, retrieving encrypted images according to the query trapdoor provided by the query user, returning a Top-k encrypted image structurally similar to the query image, and then embedding specific watermarks generated by the certification authority for each user into a reserved watermark embedding space of a Top-k encrypted image; and   S4, performing, by the certification authority, both user authentication and retrieval result correctness verification tasks, realizing the user authentication by extracting and comparing specific watermark information in the suspicious image, and realizing the retrieval result correctness verification by reconstructing the Merkle tree and comparing the root hash stored by the image owner.   
     
     
         2 . The remote sensing image security retrieval method in cloud environment according to  claim 1 , wherein in the S1, a construction method of the encrypted searchable index is as follows:
 S21, extracting a feature vector of each image by using a pre-trained CNN model;   S22, converting high-dimensional feature vectors into low-dimensional binary codes according to spectral hashing with spectral rotation, and performing K-means++ clustering; and   S23, expanding the feature vector after dimensionality reduction and clustering, and splitting into two parts, and encrypting the two parts obtained by splitting respectively according to two random matrices.   
     
     
         3 . The remote sensing image security retrieval method in cloud environment according to  claim 1 , wherein in the S1 and the S3, the image set encryption and watermark embedding methods are as follows:
 S31, performing prediction error marking and pixel rearrangement to create watermark embedding space;   S32, generating a pseudo-random matrix by a pseudo-random generator, binarizing the current pixel, and performing bitwise XOR encryption; and   S33, extracting starting position information for data embedding in each image, and then embedding the watermark in the space vacated by the encrypted image in the form of pixel replacement after the starting position.   
     
     
         4 . The remote sensing image security retrieval method in cloud environment according to  claim 2 , wherein in the S2, the query trapdoor generation is to expand the query vector and split the query vector into two parts, and then encrypting the two parts according to two random matrices, wherein the two random matrices and the two random matrices in S23 are reciprocal quantities. 
     
     
         5 . The remote sensing image security retrieval method in cloud environment according to  claim 1 , wherein in the S4, the authenticity verification of the retrieval result by the certification authority is as follows:
 S51, returning, by the certification authority, all adjacent nodes along the path from the root node to the leaf node for the image to be verified as evidence according to the Merkle tree of the image set; and   S52, constructing a new Merkle tree according to the image to be verified and its evidence, and comparing the root hash values of the new Merkle tree with the root hash values of the original Merkle tree.   
     
     
         6 . The remote sensing image security retrieval method in cloud environment according to  claim 1 , wherein in the S4, the tracking of the distribution data by the certification authority is as follows:
 S61, identifying, by the image owner, suspicious images and transmitting both duplicate copies and original images to the certification authority; and   S62, locating, by the certification authority, the reserved watermark-embedded space by extracting the starting position marker, and recovering the embedded watermark information.   
     
     
         7 . A remote sensing image security retrieval system in cloud environment, comprising a retrieval index generation and encryption module, an image encryption and watermark embedding module, a query vector generation and encryption module, a retrieval module, a retrieval result authenticity verification module and a data distribution legitimacy verification module; the remote sensing image security retrieval system in cloud environment executes the remote sensing image security retrieval method in cloud environment according to  claim 1 ;
 wherein the retrieval index generation and encryption module is configured for the image owner to extract low-dimensional image features according to a CNN model and spectral hashing with spectral rotation, and encrypt the feature vector by using ASPE to form the encrypted searchable index, and send the encrypted searchable index to the cloud server; the image encryption and watermark embedding module is configured for the image owner to encrypt the image after the image owner rearranges the reserved space according to the prediction error, and the cloud server embeds the watermark containing the identity information of the query user for data tracing; the query vector generation and encryption module is configured for the query user to extract image features according to the CNN model and spectral hashing with spectral rotation, and encrypt the query vector by using ASPE to form an encrypted query vector and send the encrypted query vector to the cloud server; the retrieval module is configured to calculate the Euclidean distance between the encrypted feature vector and the encrypted query vector, and return Top-k encrypted images in ascending order; the retrieval result authenticity verification module is configured to prevent malicious cloud servers and external attackers from forging and tampering with image content by constructing the Merkle tree; the data distribution legitimacy verification module is configured to track the distributed data through digital watermarking and prevent malicious query users from leaking the retrieval result.